Sie sind auf Seite 1von 6


THEJOURNAL OF ~MMUNOLOCY Voi. 149. 2709-2714. No. 8.October 15. 1992
Copyrlght 0 1992 by The Amerlcan Assoclatlon of Immunologlsts Printed in U.S.A.




From the ‘Departmentof Surgery, Universityof Minnesota, Mtnneapolis,M N 55455; and ‘From the Jackson Laboratory,
Bar Harbor, ME 04609

The potential of the immune system to inhibit or it is notclear that cancer is routinely eliminated by
stimulate tumor growth is a vivid example of the “immunosurveillance”as originally envisaged by Ehrlich
*two-edged sword” nature of immune responses. Our ( 1 ) and developed by Burnet (8).That is, a picture of the
results provide evidencethat thisdual capacity can immune system as serving a purely inhibitory role in
be attributed, in part, to thedual pathways of argi- cancer had to be rethought whenit was discovered during
nine metabolism exhibited by intratumor macro- the 1970’s that immunodeficient hosts did not exhibit an
phages. Specifically, i.p. tumor rejection in P815- increased incidence of cancer (with the minor exception
preimmunized miceis accompanied by an upshift in of lymphoid malignancies) (5, 9, 10). Also during this
intratumor macrophage arginine metabolism to the
nitric oxide (NO) synthase pathway that yields ci- period, Prehn and co-workers (11, 12) advancedthe the-
trulline and NO. A rapid and marked local increase ory that the immunesystem could actively stimulate
in IFN-y (both mRNA and protein) in preimmunized cancer. Although the “immunostimulation” theory has
mice during tumor rejection suggests that this cy- received little attention by most investigators in their
tokine plays a role in up-regulating nitric oxide pro- pursuit of anticancer defenses, sufficient evidence has
duction in vivo. Unliketumor rejection, progressive accumulated to conclude that cancergrowth can be stim-
i.p. P815 tumor growth in naive mice is associated ulated a s well as inhibited by the immune system (13-
with a marked decline in the production of citrul- 15).In particular, macrophages have been demonstrated
line/NO by intratumor macrophages. Examination toexhibit this dualcapacityunderdifferentcircum-
of macrophage arginine metabolism via arginase stances (16-21).
revealed a pattern opposite that of NO synthase. Recent results in this laboratory (22-24) and others
The local production of ornithine/urea markedly (25-28)havedemonstrated thatthe pathway macro-
increases during progressive tumor growth whereas phages use to metabolize arginine canpromote or inhibit
arginase activity decreases during tumor rejection. their functions, and thatof cells with which they com-
Inasmuch as nitric oxide inhibits tumor cell repli- municate. This knowledge combined with the aforemen-
cation whereas ornithine is the precursor of poly-
amines required for cell replication, these results tioned background in tumor immunology led us to pos-
are consistent with the conclusion that thepathway tulate that macrophage arginine metabolism at the site
macrophages use to metabolize arginine can influ- of atumor could influencewhethertumor growth is
ence the type of host immune responses against inhibited or stimulated. It was proposed that macrophage
cancer and otherconditions. arginine metabolism in the tumor bed via the NO3 syn-
thase pathway would favor tumor inhibition becauseNO
is tumoristatic/tumoricidal(26).If instead, NO production
Despite intensive and prolonged researchefforts in was down- regulated, tumor growth might be allowed or
tumor immunology dating back to the turn of the century even stimulatedbecausemacrophages are aprimary
when Ehrlich(1) and others (2) proposed first an antican- source of growth factors that promote proliferation of a
cer role for the immune system, there are still no con- variety of cells such as those involved in wound healing
sistently effective protocols for the immunotherapy of
and regeneration (29, 30). Further, if the predominant
human cancer. Clearly the immune system has the po-
pathway of arginine metabolism by intratumor macro-
tential to inhibit cancer. Indeed, several different potent
phages wasvia arginase (yieldingornithine and urea) (31)
antitumor mechanisms have been identified in animal
models (3.4, reviewed in 5-7).At the sametime, although rather than NO synthase then tumor growth might be
anticancer potential can be demonstrated experimentally further promoted because ornithine is the precursor of
polyamines that arerequired for cell replication (32, 33).
This report examines these hypotheses. A model was
Received for publication April 6, 1992.
Accepted for publication August3, 1992. used in which tumor growth or rejection occurs in the
The costs of publication of this article were defrayed in part by the peritoneum. This model facilitates analyses of the intra-
payment of page charges. This article must therefore be hereby marked tumor cellular infiltrate and local extracellular environ-
aduertfsement in accordance with 18 U.S.C. Section 1734 solely to indi-
cate this fact. ment (peritoneal fluid).It will be shown thattumor rejec-
This work was supported by National Institutes of Health Grant tion or growth is associated with local up regulation of
GM3224 and CA27523(RE). Certain of the results in this paper were
presented in preliminary form at the 2nd InternationalMeeting on Nitric the N O synthase or arginasepathway, respectively.
Oxide in London, October1991. These and other data suggest that intratumor macro-
‘Address correspondence and reprint requests to Charles D. Mills,
Department of Surgery, University of Minnesota, Box 120 University of Abbreviations used in this paper: NO, nitric oxide: PEC. peritoneal
Minnesota Hospitals and Clinics, Minneapolis,MN 55455. exudate cells.

phage arginine metabolism plays a role in determining supernatants andperitoneal fluids.

what course cancer takes. Northern hybridization. RNA was extracted from cells by a mod-
ification of the guanidine hydrochloride technique described fully
elsewhere (37.38). TheIFN--, cDNAprobe was obtained from DNAX.
MATERIALS AND METHODS (Palo Alto, CA). Allprobes were labeled
with [32Pl dCTP by the method
of Feinberg and Vogelstein (39) to a sp. act. of =lo9 dpm/pg cDNA.
Animals. C57B1/6 X DBA/2 (B6D2F1) female mice 8-wk-old were In all experiments, RNA was blotted onto ZetaProbe nylon mem-
purchased from Taconic Farms, Germantown. NY. Mice were used branes (Bio-Rad Laboratories, Richmond, CA) and hybridized with
in experiments when they were 10- to14-wk-old. They werefree of = 1O6 cpm/ml hybridization mix. Blots were exposed to XAR-5 x-ray
common murine pathogens as assessed by routine testing by the film (Kodak) between 2 Du Pont Cronex Lightening Plus (Du Pont.
supplier and by the University of Minnesota animal care facility. Wilmington. DE) screens for variable times a t -70°C. Densitometry
Mice were killed with carbon dioxide gas. was carried out on the autoradiograms using the Bio-Rad model620
Tumors. The P815 mastocytoma employed has been described in Video Densitometer (Bio-Rad).Data were normalized relative to the
detail previously (3,34. 35).Briefly, the tumor waspassaged weekly/3-actin signal.
in vivo in irradiated (750R) B6D2F1 mice. Mice were specifically
immunized against P815 by axillary intradermal injection of 2.5 x RESULTS
lo6 tumor cells mixed with 50 pg Corynebacterium paruum (Bur-
roughs Wellcome Ltd.. London, England). By 21 days postimmuni- Properties of tumor model used to determine intra-
zation complete tumor regression has occurrred. At this time. im-
munized or age-matched naive mice were challenged with 5 x l o 5 tumor arginine metabolism.Mice exhibiting specific im-
P815 tumor cells in the peritoneum in 0.5 ml of PBS. For in vitro munity to the P815 wereproduced as in preceding pub-
use. P815. L5178Y. and YAC tumor cells were grown inRPMI 1640 lications (3,34, 35). Specifically, B6D2Fl mice receiving
supplemented with 10% FCS and antibiotics (penicillin/streptomy-
cin). New stocks of tumors were used in vivo and in vitro every few
intradermal injections with P815 tumor cells mixed with
months from cells kept frozen in liquid nitrogen. Corynebacterium paruum reject the tumor in 12 to 1 5
PEC and peritonealfluid.PEC and peritoneal fluid werecollected days and thenexhibit long-lived memory immunity. For
simultaneously by the injection of 0.5 ml PBS containing antibiotics our experiments, immunized (21 days postimmunization)
followed by immediate recovery of available fluid/cells. Injecting a
relatively small volume reduces the dilution of peritoneal fluid and or naive mice were challenged with 5 X 1O5 P8 15 tumor
yet allows more than half of the PEC to be recovered. Recovery of cells in theperitoneum and the intratumorcellular infil-
fluid was quite consistent averaging 0.42 f 0.04 ml in normal mice. trate (PEC) and the local extracellular environment (per-
The volume of the peritoneal fluidwas determined using radioisotope
dilution. Mice received injections in the peritoneum with 0.02 PCi
itoneal fluid) were analyzed during tumor growth or re-
[51Cr]in 0.5 mlPBS and then peritonealfluid was immediately jection. It can be seen first Figure in 1A that exponential
recovered. A comparison of the radioactive counts injected and re- growth of tumor occurs in naive mice such that by
9 days
covered allowed the peritoneal fluid volume to be calculated. postimplantation >lo8 P815 tumor cells are present in
Cell culture. Details of cell culture have been previously reported
(24).PEC were washed and resuspended in complete culture media the peritoneal cavity (line graphs).In contrast, in preim-
consisting ofRPMl 1640 supplemented with 10% FCS (Sterile Sys- munized mice thereare less than1O5 tumor cells present
tems, Logan, UT), 10 mM morpholinopropane sulfonic acid, 5 X by 5 days after tumor implantation. A s previously ob-
M 2-ME, and antibiotics. Unless indicated, these and other tissue
culture reagents were purchased from GIBCO/BRL. Gaithersbur
served at other anatomical sites (34), tumorrejection in
MD. Cells were cultured in 96-well flat-bottomedplates a t 1.5 X 10 / the peritoneum of preimmunized miceis accompanied by
ml in a volume of 0.2 ml for 20 h a t 37°C in 7% CO, in air. Cell the appearance of robust cytolytic activity (Fig. lA, his-
supernatants for NO, or amino acid analysis were collected and tobars) mediated by tumor-specific T lymphocytes of the
stored a t -80°C. Macrophage-rich cultures were obtained by a 2-h
incubation of 3 X lo5 PEC in 0.1 ml followed by repeated washing Lyt-2+ (CD8),L3T4- (CD4)phenotype (Fig. 1B).A s shown
(22). ThePEC that did not adhere to the wells were placed in other previously, T lymphocytes are known to be required for
wells and referred to as nonadherent. tumor rejection and immunity is tumor specific (6, 35).
[51Cr] release assay. The assay was the same as that used previ-
ously (34). Briefly, P815 or other tumor cells were harvested during In contrast, no significant cytolytic activity was detecta-
log-phase growth and lo6 cells were labeled in 0.3 ml of media with ble during progressive P8 15 tumor growth in naive mice.
100 pCi of Naz[51Cr]04 (CJS. 11, Amersham Corp, Arlington Heights, Dgferent pathwaysof intratumor leukocyte arginine
IL) for 1 h a t 37°C. PEC and [51Cr]-labeledtumor target cells (5 X lo3) metabolismpredominateduring tumorrejection or
were plated in round-bottomed 96-well plates in quadruplicate in a
total volume of 0.2 ml and the plate was centrifuged to initiate the growth. Examination of arginine metabolism at the site
assay. After 4h at 37°C. 0.05 ml of medium was removed and of tumor rejection or progressive tumor growthrevealed
counted (LKB Instruments, Inc.. Gaithersburg. MD). “Total” (5’Cr] fundamental differences. First, it can be seen that leu-
release was thatreleased by 0.5% Triton-X. Total releasewas >97%
and spontaneousrelease was 6 to 10%of the total depending on the kocytes recovered from preimmunized mice 1 day after
tumor. The percentage of specific [51Cr]release was calculated as tumor implantation demonstrate increased metabolism
follows: (experimental cpm - spontaneous cpm/total cpm - sponta- of arginine to citrulline and NO (measured as sodium
neous cpm)X 100. nitrite) (Fig. 2, A and B). Also, the up-regulation of the
Antibody treatment. T lymphocyte subsets were depleted as pre-
viously described (35). PEC (2 x 107/ml)were incubated for 30 min NO synthase pathway is maintained for at least 1 wk
a t 4°C in a 1/5 dilution of a hybridoma culture supernatant. Anti- after tumor implantation.In contrast, during progressive
L3T4 (GK1.5) and anti Lyt-2 (TIB-150) hybridomas were obtained tumorgrowthinnaivemice, local leukocytearginine
from theAmerican TypeCulture Collection (Rockville, MD). Low-tox
rabbit complement (AccurateChemical and Scientific Co., Westbury, metabolism to citrulline andNO declines below that ob-
NY) was added and cells were incubated for a n additional 60 min at served with resident normal peritoneal exudate leuko-
37°C. cytes. Second, precisely the opposite pattern is observed
Nitrite determination. Nitrite in cell-free culture supernatants when arginine metabolism via the arginase pathway is
was measured as described by Ding et al. (36).Briefly, 0.05 ml of
test fluid was mixed with 0.05 ml of Griess reagent in quadruplicate evaluated. Thatis, it can be seen that leukocyte arginine
and absorbance read at 550 nm using a n automatic plate reader. metabolism to ornithine and urea markedly increases
Nitrite concentration wascalculated from a NaNO, standard curve. during progressive tumor growth, whereas during tumor
Amino acid analysis. Amino acids were measured as previously
described (22) using a Dionex BioLC Amino Acid Analyzer System rejection the arginase pathway in leukocytes is decreased
(Dionex Corp.. Sunnyvale. CA) with lithium eluents and norleucine below normal (Fig. 2, C and D).
as a n internal standard. Variationbetweenreplicate HPLC runs Intratumor shifts in NO synthase and arginase are
routinely averaged less than 5%.
IFN-7. A solid phase. two-site “sandwich immunoassay”(Arngen, primarily attributable to macrophages. To confirm that
ThousandOaks, CA) was used tomeasure IFN-7 in cell culture the changes in leukocyte arginine metabolism observed
Figure 1. Properties of tumor model
used to analyze intratumor arginine me-
tabolism. A. Kinetics of tumor rejection or
progressive tumor growth (line graphsin
background]inimmune or naivemice.
Cytolytic activity (histobars in fore-
ground] of PEC from immunized or naive
mice on the days indicated following im-
plantation of 5 x IO5 P815 tumor cells in
peritoneum.Unfractionated PEC
pooled from five mice tested at a 50:1 E:T
ratio in quadruplicate in a 4-hI5’Cr] re-
lease assay. B,Phenotype and specificity
of cytolytic cells from immunized mice 4
days after P8 15 tumor implantation ( 10:1
E:Tratio). Cytolytic phenotype(Lyt-2+
L3T4-) determined by treating PEC with
the indicated mAb or a controlculture
specificity of the PEC for the P815 tumor
was determined by incubating PEC with
[51Cr]-labeled L5178Y (another immuno-
genictumor of DBA/2 origin) or [”Crl-
labeled YAC- 1, a tumor sensitive to lysis
by NK cells.Representativedatafrom
three experiments.

W Immune 1200 1j
0 Naive 1000

L1 1,
60 Media

FLgure 2. Changesinintratumorleukocyte 20
arginine metabolism during tumor rejection or
progressive tumorgrowth. PEC were collected
from groups of five immunized or naive mice 0
after P815 tumor implantation on the days indi- 3
cated. PEC were culturedat 1.5 x 10B/ml(0.2 ml) 1 3 5 7 1 5 7
in quadruplicate in 96-well microtiter plates for 30
20 h . Cell-free supernatant was collected and
stored at -80°C until analyzed. “Normal” = resi- 0
dentperitonealexudateleukocytes of normal
mice. “Media”= cell free media cultured for20 h.
Citrulline. ornithine, and urea were analyzedby
-5 2ooo

HPLC: nitrite (NO2)determined using the Griess
reagent. The same pattern of intratumor leuko-
cyte arglnine metabolism was observedIn three
nts. similar 2! 10

1 3 5 7 1 5 7


in the preceding section could be attributed to macro- onstrated elevated ornithine production as compared
phages, PEC were fractionated. Peritoneal exudate leu- with resident PEC from normal mice. Again, the majority
kocytes were incubated for 2 h at 37°C: leukocytes re- of leukocyte ornithine production from naive tumor bear-
maining adherent afterthorough washingare “adherent” ers or normal mice is found in the adherent fraction.
and the rest are “nonadherent.” Adherent cells comprised Although a percentage (approximately 50%)of the PEC
on average 54% of immune PEC and 45% of naive PEC. in naive mice at this time are tumor cells, P815 tumor
More than 93%of adherent immune or naive PEC were cells are nonadherent and produce little ornithine. Spe-
macrophages as assessed by morphology or nonspecific cifically, a culture supernatantof P815 tumor cells con-
tained 36 gM ornithine after 20 h incubation at 1.5 X
esterase staining. Themajority of remaining cells exhib-
106/ml (the maximum cell density if all PEC were tumor
ited lymphocyte-like morphology. It can be seen first in cells]. Together, these data indicate that macrophages
Figure 3A that, again, local NO production by peritoneal are primarily responsible for the increased local citrul-
exudate leukocytes from immunized mice is elevated 5 line/NO or ornithine/ureaproduction during tumorrejec-
days aftertumor implantation. More importantly, Figure tion orprogressive tumor growth,respectively.
3A shows that virtually all of the N O (measured a s NO2) NO synthase or arginase pathway predominates in
produced can be attributed to the adherent cell fraction. local extracellular tumor environment during rejection
In Figure 3B it can be seen that PEC recovered instead or growth, respectively. Direct examination of arginine
from naive mice 5 days after tumor implantation dem- metabolism at the site of tumor rejection or progressive

301 A
W Immune

Figure 3. Evidence that macrophages are
primarily responsible for changes in intratu-
mor leukocyte arginine metabolism. PEC
were collected from groupsof five immune or
five naive mice 5 days after P815 tumorim- 10
plantation.Unfractionated PEC, adherent
PEC. or nonadherent PEC were cultured over-
night a s in Figure 2.

Non-AdherAdherent Un-Fract.
Adherent Un-Fract.
0 h
tumor growth (peritoneal fluid)revealed the same char- to be maximal 1 day after tumor implantation and to
acteristic differences observed with intratumor leuko- disappear by day 5. In support of these findings, Figure
cytes. That is, the local concentration of citrulline is 5C shows that leukocyte mRNA coding for IFN-7 is abun-
elevated in peritoneal fluid in immunized mice within 1 dant early after tumor challenge and rapidly declines
day of tumor implantation and remains somewhat ele- indicating a close correspondence between gene tran-
vated during tumor rejection (Fig. 4A).Inasmuch as ci- scription and translation/secretion. Finally, because lo-
trulline andNO are produced from argininein equimolar cal production of citrulline/NO remains elevated during
quantites by NO synthase (25), the concentration of this tumor rejection for at least1 wk, the data suggest that if
amino acid shouldreflect NO synthase activity. In IFN-y is responsible for increasedNO production in vivo,
marked contrast to these findings, it can be seen that then it is either required only transiently or that other
during progressive tumor growth, the intratumor citrul- signals play a role as well.
line concentration declines to levels significantly below In contrast to our knowledge of factors such as IFN-y
that found inperitoneal fluid from normal mice. Exami- that “activate” macrophages, less is known concerning
nation of arginine metabolism via arginase revealed pre- the signals that down-regulate local macrophage arginine
cisely the opposite pattern to that of N O synthase. Spe- metabolism to citrulline/NOand increase the production
cifically, there wasa progressive rise in thelocal concen- of ornithine/urea during progressive tumor growth. How-
tration of ornithine in peritoneal fluid during progressive ever, some informed speculation maybe useful.
tumor growth, but not during tumor rejection (Fig. 48). A tumor could regulate intratumor macrophage argi-
In contrast, thelocal concentration of ornithine is below nine metabolism directly or indirectly through other leu-
normal during specific tumor rejection. kocytes or another organ system. Regarding directregu-
Regulation of local arginine metabolism during tumorlation, tumors have been reported to produce a variety of
rejection or growth. Our results indicate that tumor re- products that could affect macrophage arginine metabo-
jection and progressive tumor growth are associated with lism. In particular, Nathan and co-workers(40, 41)have
characteristic and opposite shifts in macrophage argi- isolated a protein from several murine tumors that down-
nine metabolism. A s for the signals that regulate these regulates theNO synthase pathway in macrophages. One
changes in arginine metabolism, IFN-y has been reported tumor that secretes this protein (macrophage deactivat-
(36) to up-regulate the NO synthase pathway. Evidence ing factor) is the P815 mastocytoma used in our study.
consistent with the ability of IFN-y IFN to up-regulate Thus, it seems reasonable to postulate that such a factor
this pathway is shown in Figure 5A. It can be seen that plays a role in down-regulatingthe NO synthase pathway
the local extracellular concentration (peritoneal fluid) of during progressive P815 tumor growth. Another factor
IFN-y is markedly increased in immunized mice, but not that seems likely to play a regulatory role in intratumor
naive mice, after tumor implantation. Evidence that in- macrophage arginine metabolism is TGF-P. This cytokine
tratumor leukocytes are responsible forlocal the increase is produced by some tumors and has been shown to down-
in IFN-y is shown in Figure 5B where it can be seen that regulate NO production in macrophages in vitro (41).CSF
leukocytes recovered fromthe siteof tumor rejection, but includinggranulocyte-macrophage CSF arealso pro-
not from the site of tumor growth, secreted markedly duced by certain tumors (42) and these cytokines have
elevated levelsof IFN-7. Leukocyte secretion of IFN-r a n d been demonstrated to inhibit activation-associated func-
its concentration in peritoneal fluid were both observed tions of macrophages (43).Although CSF have not been


0 Naive

Normal 60
FLgure 4. Changes in arginine metabolism in r

the local extracellular environment during tumor

regression or progressive tumor growth. Perito-
neal fluid was collected and pooled from groups
of five immunizedor naive mice on the indicated =
days after P815 tumor Implantation. Citrulline
and ornithine concentrations were determined 0 20
using HPLC. A virtuallyidenticalpattern of
changes in citrulline and ornithine in peritoneal
fluid wasobserved in another experiment of the 3 5 0 5 9
same design. Normal peritoneal fluid represents
1 7 9 1 3
the mean & SD of five mice.

Days Post Tumor Implantation

AND 2713
80 tion of macrophages stimulatedby IFN-7 (48,49). Finally,
Immune other organ systems could contribute to indirect regula-
0 Naive (All c 1 U/ml) tion of arginine metabolism in macrophages. In particu-
. E
60 lar,participationbythepituitary-adrenalaxisseems
likely because circulating steroids are sometimes elevated
in cancer(50)and steroids have been shown (51) to down-
regulate NO synthase activity.
E Regarding the regulation of macrophage arginine me-
W 20 tabolism, the relationship between NO production and
ornithine/urea production via the arginase pathway is
largely unknown. It isknown (52)that bothenzyme
1 3 5 7 9 activities can increase simultaneously in macrophages
becauseactivatedmurinemacrophages producemore
B citrulline/NO and ornithine/urea than resident macro-
phages. In the tumormodel used in this study, citrulline/
- 8 NO production and ornithine/urea productionby macro-
phages were inversely related. Relatedly, it is known that
6 blocking the NO synthase pathway with N-G-monome-
2 thyl-L-arginine increases ornithine production by murine
E 4 macrophages (52). An increase in ornithine production
5 as a consequence of inhibition ofNO production would
be expected because the K , of arginase is higher than
that ofNO synthase (51).However, whether down-regu-
lation of NO production necessarily increases ornithine
1 3 5 7 9
production is not known. Ongoing experiments are de-
signed to more clearly delineate how these two pathways
of arginine metabolism are regulated in macrophages.
Our results provide evidence that intratumor macro-
phage arginine metabolism is a molecular explanation
for the dual ability of the immune system to inhibit or
stimulate tumor growth.It is proposed that arginine me-
tabolism in the tumor bed yielding citrulline and NO
favors tumor rejection, whereas production of ornithine
and ureacould promote tumor growth. Inasmuch as mac-
rophages are the primary intratumor sourceofNO syn-
thase, their presence in a tumor cocld be considered a
favorableprognosis. However, if intratumormacro-
phages are expressing the arginase pathway then their
1 3 5 7 9
presence could be unfavorable. It should be added that
although ornithine-derived polyamines can stimulate tu-
DAYS POST TUMOR IMPLANTATION mor growth,we do not have direct evidence of stimulation
Figure 5. Local elevation in IFN-y during tumor rejection. Peritoneal of the P815 tumor cells used in this study. Although it is
fluid and PEC were collected from immunized or naive mice after tumor
implantation as in preceding figures. A . IFN-y concentration inperitoneal a cruel paradox that macrophages can promote as well
fluid. E. Leukocyte secretion of IFN-a into supernatant after overnight as inhibit tumor growth, this dual capacity is in keeping
culture as in Figure 2. C . LevelofmRNA determined by densitometry
tracing of Northern blots. IFN-ymRNA normalized for B-actin mRNA with the multiple roles that these cells physiology. play in
content of each sample. No detectable IFN-y or IFN-y mRNA was present In conclusion, our results suggest that the ability mod- to
in normal mice. ulate immune responses against cancer and other con-
ditions can depend, in part, on stimulating the appropri-
shown to inhibitN O production, per se, it seems reason- ate pathwayof arginine metabolism in macrophages.
able to postulate that they might.In particular, CSF can
promote macrophage proliferation (44) which seems un- Acknowledgments. The authors gratefully acknowl-
likely during precocious NO production because macro- edge the excellent technical assistance of Lisa A. Frie,
phages themselves are susceptible to the toxic effects of Sonya Kamdar for carrying out the Northern hybridiza-
their NO (22, 23, 27). tion analysis, and the helpof Patti Jukich in the prepa-
It is likely that tumors also regulate macrophage argi- ration of the manuscript.
nine metabolism indirectly. In particular, the immune
system itself is a likely source of indirect regulation by REFERENCES
tumor cells because lymphocytes produce TGF-P and CSF 1 . Ehrlich, P. 1909. Uber den jetzigen standder karzinomforschung. In
(45.46). In this connection, the type of indirect regulation The Collected Papers of Paul Ehrlich. Vol. 2. F. Himmelweit. ed.
that macrophages receive may depend on the type of T Pergamon Press, London, p. 550.
2. Tyzzer, E. E. 1916. Tumor immunity. J. Cancer Res. 1:125.
lymphocyteactivated. For example, Th 1 lymphocytes 3. Mills, C. D.. R. J. North, and E. S. Dye. 1981. Mechanisms of anti-
produce IFN-7, and thus could increase NO production tumor action of Corynebacterium parvum. 11. Potentiated cytolytic T
cell-responseandits tumor-induced suppression. J . Exp. Med.
(47).However, Th2 lymphocytes produce IL-4 and IL-10 154:621.
both of which have been shown toblock the NO produc- 4. Evans. R. 1983. Combination therapy by using cyclophosphamide
and tumor-sensitized lymphocytes:a possible mechanismof action. 30. Freundlich, B., J. S. Bowaliski, E. Neilson. and S. A. Jimenez.
J . rrnrnunol. 130:2511. 1986. Regulation of fibroblast proliferation and collagen synthesis
5. Gorelick, E. 1983. Concomitant tumor immunity andthe resistance by cytokines. Imrnunol. Today 7:303.
to a second tumor challenge.Adu. Cancer Res.39:71. 31. Schneider. E., and M. Dy. 1985. Therole of arginase in the immune
6. North, R. J. 1984. The murine anti-tumor response and its thera- response. Irnrnunol. Today 6: 136.
peutic manipulation.Adu. Irnrnunol. 95:89. 32. Williams-Ashman, H. G., and 2. N. Canellakis. 1979. Polyamines in
7. Greenberg, P. D. 1991. Adoptive T cell therapy of tumors: mecha- mammalian biology and medicine. Perspect.Biol. Med. 22:42I.
nisms operative in the recognition and elimination of tumor cells. 33. Pegg, A. E. 1988. Polyamine metabolismanditsimportancein
Adu. Irnmun. 49: 281. neoplastic growth and as a target for chemotherapy. Cancer Res.
8. Burnet, F. M. 1971. Immunological surveillanceinneoplasia. 48:759.
Transpl. Reu. 7: 3. 34. Mills,C.D.. and R. J. North. 1983. Expression of passively trans-
9. Stutman. 0.1975. lmmunodepressionand malignancy. Adu. Cancer ferred immunity against a n established tumor depends on generation
Res. 22: 261. of cytolytic T cells in recipient. Inhibitionby suppressor T cells. J .
10. Grossman, 2..and R. B. Herberman. 1986. “Immune surveillance” Exp. Med. 157: 1448.
without immunogenicity.Imrnunol. Today 7: 128. 35. Mills, C. D., and R. J. North. 1985. Ly-1+2-suppressor cells T inhibit
11. Prehn, R. T., and M. A. Lappe. 1971. An immunostimulation theory the expressionof passively transferred anti-tumor immunityby sup-
of tumor development. Transplant. Rev. 7:26. pressing the generation of cytolytic T cells. Transplantation 39:202.
12. Prehn, R. T. 1972. The immune reaction a s a stimulator of tumor 36. Ding, A. H., C. F. Nathan, and D. J. Stuehr. 1988. Release of reactive
growth. Science 176:170. nitrogenintermediatesandreactive oxygen intermediatesfrom
13. Murasko. D. M., and R. T. Prehn. 1983. The ability of immune mouse peritoneal macrophages: comparison of activating cytokines
reactivity to potentiate tumor growth. In Imrnunobiology of Trans- and evidence for independent production.J . Immunol. 141:2407.
plantation. Cancer and Pregnancy. P. K. Ray, ed. Pergamon Press, 37. Evans. R., and S.J. Kamdar. 1990. Stability ofRNA isolated from
New York. p. 148. macrophages depends on the removal of a n RNA-degrading activity
14. Yamagishi, H., N. R. Pellis, and B. D. Kahan. 1983. Specific and early in the extraction procedure. BioTech. 8:357.
nonspecific immunologic tumor growth facilitation. In Irnrnunobiol- 38. Kamdar. S.J., and R. Evans. 1992. Modifications of the guanidine
ogy of Transplantation, Cancer and Pregnancy. P. K. Ray,ed. hydrochloride procedure for the extractionof RNA: isolation from a
Pergamon Press, New York. p. 179. variety of tissues and adherent/nonadherent cell types. BioTech.
15. Fidler, I. J. 1990. Critical factors in the biology of human cancer 12:632.
metastasis: twenty-eighth G. H. A. Clowes Memorial Award Lecture. 39. Feinberg. A. P., and B. Vogelstein. 1983. A technique for radiola-
Cancer Res. 50:6130. beling DNA restriction endonuclease fragments to high specific ac-
16. Carrel, A., and A. H. Ebeling. 1926. The fundamental propertiesof tivity. Anal. Biochem. 132:6.
the fibroblast and the macrophage.11. The macrophage.J . Exp. Med. 40. Srimad. S..and C. Nathan. 1990. Purification of macrophage deac-
44:285. tivating factor. J . Exp. Med. 171~1347.
17. Nathan, C.F., and W. D. Terry. 1975. Differential stimulation of 41. Ding, A., C. F. Nathan, J. Graycar. R. Derynck, D. J. Stuehr, and
murinelymphomagrowthinvitro by normal and BCG-activated S. Srimal. 1990. Macrophage deactivating factor and transforming
macrophages. J . Exp. Med. 142887. growth factors-81, -62, and - 83 inhibit induction of macrophage
18. Hibbs, J. B.. Jr. 1976. The macrophage a s a tumoricidal effector nitrogen oxide synthesis by IFN-7 J . Irnrnunol. 145:940.
cell: a review of in vivo and in vitro studies on the mechanism of the 42. Takeda. K.. K. Hatakeyama. Y. Tsuchiya, H. Rikiishi. and K. Ku-
activated macrophage nonspecificcytotoxic reaction. In The Macro- magai. 1991. A correlation between GM-CSF gene expression and
phage in Neoplasia. M. A. Fink, ed. Academic Press, New York. p. metastases in murine tumors.Int. J . Cancer 47:413.
83. 43. Sotomayor, E. M., Y.-X. Fu, M. Lopez-Cepero, L. Herbert, J. J.
19. Evans, R. 1978. Macrophage requirement for growth of a murine Jimenez, C. Albarracen, and D. M. Lopez. 1991. Role of tumor-
fibrosarcoma. Br. J . Cancer 37: 1086. derived cytokines onthe immune system of mice bearing a mammary
20. McBride, W. H. 1986. Phenotype and functions of intratumor mac- adenocarcinoma: Down-regulation of macrophage-mediated cytotox-
rophages. Biochim. Biophys. Acta 865:27. icity by tumor-derived granulocyte-macrophage colony-stimulating
21. Buick, R. N.. S. E. Fry. and S. E. Salmon. 1980. Effect of host-cell factor. J . Irnrnunol. 147:2816.
interactions on clonogenic carcinoma cells in human malignant ef- 44. Nakata. K.. K. S.Akagawa, M. Fukayama, Y.Hayashi. M. Kadok-
fusions. Br. J . Cancer 41:695. ura. and T. Tokunaga. 1991. Granulocyte-macrophage colony-stim-
22. Albina, J. E., M. D. Caldwell, W. L. Henry, and C. D. Mills. 1989. ulating factor promotes the proliferation of human alveolar macro-
Regulation of macrophagefunctions by L-arginine. J . Exp. Med. phages invitro. J . Irnrnunol. 147: 1266.
45. Silva, J. S..D. R. Twardzik, and S. G. Reed. 1991. Regulation of
23. Albina, J. E., C. D. Mills, W. L. Henry, and M. D. Caldwell. 1989. Trypanosoma cruzf infections in vitro and invivo by transforming
growth factor 0 (TGF-6).J . Exp. Med. 174~539.
Regulation of macrophage physiology by L-arginine. J . Irnrnunol.
46. Kelso, A., and D. Metcalf. 1990. T-lymphocytederived colony-stim-
143:3641. ulating-factors. Adu. Irnrnunol. 48:69.
24. Mills, C.D. 1991. Molecular basis of “suppressor”macrophages: 47. Mosmann, T.R., and R. C. Coffman. 1987. Two types of mouse
argininemetabolism via the nitric oxide synthetase pathway. J . helper T-cell clone. Irnrnunol. Today 8:223.
rrnrnunol. 146:2719. 48. Liew, F. Y., Y. Li. A. Severn, S.Millott, J. Schmidt, M. Salter, and
25. Hibbs, J. B., Jr., R. R. Taintor. and 2. Vavrin. 1987. Macrophage S.Moncada. 1991. A possible novel pathway of regulation by murine
cytotoxicity: role for L-arginine deiminase and imino nitrogenoxida- T helper type-2 (Th2)cells of a T h l cell activity via the modulation
tion to nitrite. Science 2355473. of the induction of nitric oxide synthase on macrophages. Eur. J .
26. Hibbs, J. B.. Jr.. 2.Vavrin, and R. R. Taintor. 1987. L-Arginine is Irnmun. 21:2489.
required for expression of the activated macrophage effector mech- 49. Gazzinelli, R. T., I. P. Oswald, S.L. James, and A. Sher. 1992. IL-
anism causing selective metabolic inhibition in target cells. J . Im- 10 inhibits parasite killing and nitrogen oxide production byIFN-
rnunol. 138:550.
Drapier, J. C., and J. B. Hibbs J r . 1988. Differentiation of murine
macrophages to express nonspecific cytotoxicity for tumor cells re-
gamma activated macrophages.J . Irnrnunol. 148: 1 792.
50. Besedovsky. H. O., A. del Ray, M. Schardt, S. Normann, J. Bau-
mann. and J. Girard. 1985. Changes in plasma hormone profiles
sults in L-arginine-dependent inhibition of mitochondrial iron-sulfur after tumor transplantation into syngeneic and allogeneic rats. Int.
enzymes in the macrophage effector cells. J . Irnrnunol. 140:2829. J . Cancer 21~209.
28 Hoffman, R. A., J. M. Langrehr, T. R. Billiar, R. D. Curran, and R. 51. Moncada, S.,R. M. J. Palmer. and E. A. Higgs. 1991. Nitric oxide:
L. Simmons. 1990. Alloantigen-induced activation of rat splenocytes physiology. pathophysiology andpharmacology.Pharrnacol. Reu.
is regulated by the oxidative metabolism of L-arginine. J . Irnrnunol. 43: 109.
145:2220. 52. Granger, D. L.. J. B. Hibbs. Jr., J. R. Perfect, and D. T. Durack.
29 Nathan, C. F. 1987.Secretoryproducts of macrophages. J . Clin. 1990. Metabolic fate of L-arginine in relation to microbiostatic ca-
Inuest. 79:319. pability of murine macrophages. J . Clin. Invest. 85:264.